[
https://issues.apache.org/jira/browse/FLINK-20926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17263323#comment-17263323
]
Q Kang commented on FLINK-20926:
--------------------------------
In my opinion, the must-have metadata definitions are as follows:
{code:java}
database STRING
table STRING
ingestion-timestamp TIMESTAMP(3) WITH LOCAL TIME ZONE // "ts" field
primary-key-columns ARRAY<STRING>
{code}
However, Maxwell still exposes extra metadata fields, such as:
{code:java}
transaction-id BIGINT // "xid" field
commit BOOLEAN
binlog-position STRING // "position" field
server-id BIGINT
thread-id BIGINT{code}
I suggest that `transaction-id`, `commit` and `binlog-position` can be kept for
further usage, while `server-id` and `thread-id` can be omitted.
[~jark] What do you think about these extra fields above?
> Allow to read metadata for maxwell-json format
> ----------------------------------------------
>
> Key: FLINK-20926
> URL: https://issues.apache.org/jira/browse/FLINK-20926
> Project: Flink
> Issue Type: Sub-task
> Components: Formats (JSON, Avro, Parquet, ORC, SequenceFile), Table
> SQL / API
> Reporter: Q Kang
> Assignee: Q Kang
> Priority: Major
>
> Canal-json and debezium-json metadata are already implemented. We still need
> metadata for maxwell-json to fill in the blanks for CDC formats.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
